The field of image quality assessment and restoration is rapidly evolving, with a focus on developing innovative methods to improve the accuracy and robustness of image processing techniques. Recent research has explored the use of deep learning-based frameworks for blind contrast quality assessment, achieving state-of-the-art performance on benchmark datasets. Additionally, there is a growing interest in learning directly on raw Bayer mosaics to avoid losses in the image-signal-processing pipeline and improve reconstruction results. Other notable trends include the use of polarimetric imaging to mitigate water-surface glare and improve semantic segmentation in aquatic environments, as well as the development of biologically inspired preprocessing techniques to enhance local contrast and color-opponency for robust semantic segmentation. Furthermore, progressive rain removal methods and extreme blind image restoration techniques are being proposed to address challenging vision tasks. Noteworthy papers include:

• A study on no-reference image contrast assessment using a customized EfficientNet-B0 model, which achieved state-of-the-art performance on benchmark datasets.
• A paper on deraining directly in the Bayer domain, which demonstrated superior reconstructions and improved results on a public benchmark.
• A work on prompt-conditioned information bottleneck for extreme blind image restoration, which proposed a novel framework for decomposing the intractable restoration process and achieved competitive results on multiple datasets.